Initial Stroke Analysis

Group 29: Helena Møller (s204585), Emma Christiansen (s204543), Casper Harreby (s204558), Frederik Holmstrup (s203540), and Laura Christiansen (s204582)

Introduction

The International Stroke Trial v. 2 data

  • Assessing efficacy and safety of antithrombotic therapy compared to control.
  • Antithrombotic agents: Aspirin and subcutaneous Heparin.

Stroke and treatments

  • Stroke is a loss of blood to the brain
  • All death causes in the data are due to possibly related blood problems

Project Aim

  • Assessing effects of antithrombotics for stroke
  • Antiplatelet treatment: Aspirin
  • Anticoagulent treatment: Heparin
  • There were no conclusive studies at this time to whether this type of treatments would work

Methods

Raw data:

  • 19435 observations with 112 variables.

01_load:

  • Data has been loaded and saved in a tsv-file.

02_clean:

  • Global syntax of NA, and Y/N for binary variables.
  • Global syntax for the use of numbers 1-8 as to describe causes of death.
  • Correction of errors made when the data was collected.
  • Removal of column determined to be irrelevant.
  • Merging of columns for easier data analysis.
  • Data was split and re-joined into one.

03_augment

  • Groups created based on different treatments, and different doses of aspirin and heparin.
# 03_augment.qmd
# Adding the different treatment groups 
data_clean_trtgroups <- data_clean_hepgroups |> 
  mutate("data_group" = case_when(
    (RXASP == "Y" & RXHEP == "L") ~ "Asp-Hep(low)",
    (RXASP == "Y" & RXHEP == "M") ~ "Asp-Hep(med)",
    (RXASP == "Y" & RXHEP == "N") ~ "Asp-Hep(No)",
    (RXASP == "N" & RXHEP == "L") ~ "Asp(No)-Hep(low)",
    (RXASP == "N" & RXHEP == "M") ~ "Asp(No)-Hep(med)",
    (RXASP == "N" & RXHEP == "N") ~ "Asp(No)-Hep(No)"
  ))

Descriptive analysis

  • Observations/patients: 19435
  • Variables: 98
  • Treatment groups: 6

Time of death after randomization up till 6 months

  • Most of the patients that die, die within the first 20 days after randomization
  • The treatment impact must therefore happen within the first period of time

Time of death after 20 days

Events during first 14 days are indicative of treatment side effects

  • Apparent relief of one cause of death has the opposite effect on another cause of death
  • Medication seems to cause fatal side effects
  • All effect sizes are relatively low

Impact of treatment on subgroups

Kaplan Meier survival curve after stroke:

  • Aspirin-Heparin(Low) treatment is the treatment with the overall highest survival rate
  • All treatments with Heparin(medium) has the lowest survival rates
  • Applying confidence intervals shows that none of the treatments are noticably away from the control treatments

Chi squared test of the treatment effect

Discussion

  • There are many conflicting data collected (e.g. date of death = 20 days, but not dead)
  • The study is very biased for Age, effect for Age should be removed
  • This study found the same as the original study, but not with an as strong effect

Conclusion

  • Aspirin has an overall very low positive effect on the patient survival
  • Herparin (Medium Dose) has an overall low negative on the patient survival
  • Based on this study, no treatment should be given